Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Planetary and Space Science 48 (2000) 1077–1086 www.elsevier.nl/locate/planspasci Filamentous fabrics in low-temperature mineral assemblages: are they fossil biomarkers? Implications for the search for a subsurface fossil record on the early Earth and Mars B.A. Hofmanna; ∗ , J.D. Farmerb; 1 a Natural b NASA History Museum, Bernastrasse 15, CH-3005 Bern, Switzerland Ames Research Center, MS-239-4, Moett Field, CA 94035-1000, USA Received 6 September 1999; accepted 15 November 1999 Abstract The subsurface has been recognized as a possible habitat for microbial life on Mars. An analogous fossil record of subsurface life is nearly missing on Earth. Here we present evidence of the widespread occurrence of such a record: tubular lamentous struc◦ tures with typical core diameters of 1–2 m were found as inclusions in minerals deposited from low-T (¡100 C) aqueous uids in subsurface environments at ¿140 localities worldwide. Filaments are frequently organized in composite structures with architectures similar to microbial mats. Filaments thickly encrusted by minerals exhibit gravity-oriented structures of macroscopic dimensions that are similar to stalactites. Environments of formation are sites of low-T water circulation in macroporous rocks such as volcanics, oxidized ores, limestone solution cavities and cavernous macrofossils. The age of occurrence ranges from Precambrian to Subrecent. We interpret the lamentous structures as permineralized and encrusted microbial laments based on the following arguments: tubular construction of laments with constant core diameters typical of microbes (1:5 ± 1:1 m), coalescence of laments to form mat-like structures, gravity draping of micron-thin laments indicating an originally exible consistency, and restriction to low-T mineral assemblages. These lamentous structures formed in subsurface environments are similar in size, morphology and construction to fossilized microbes observed in modern terrestrial and marine thermal springs. This newly recognized fossil record of subterranean microbial life opens up new perspectives in the study of the paleobiology of the terrestrial subsurface and in the exploration for fossil life on Mars. Potential host rocks on Mars include major nonsedimentary units such as volcanics and impactites. c 2000 Elsevier Science Ltd. All rights reserved. 1. Introduction Although a deep subterranean biosphere dominated by chemotrophic microbial life has been known for some time, it has only recently gained prominence (Delaney et al., 1998; Ghiorse, 1997; Gold, 1992,1999; Parkes et al., 1994; Stevens, 1997; Walter, 1996a). The recognition of biological signatures in ancient subsurface hydrothermal environments has special relevance for studies of early biosphere evolution and in the exploration for past or present life on Mars (McKay et al., 1996; Walter, 1996a; ∗ Corresponding author. Tel.: +41-31-350-7240; fax: +41-31-350-7499. E-mail address: [email protected] (B.A. Hofmann). 1 Present address: Department of Geology, Arizona State University, Box 871404, Tempe, AZ 85287, USA. Farmer and Des Marais, 1999). Previous reports of lamentous fossil microbiotas or biofabrics from subsurface environments are restricted to a few single occurrences of well-preserved laments at Warstein, Germany (Kretzschmar, 1982; Dexter-Dyer et al., 1984; Dahanayake and Krumbein, 1986), NE Scotland (Trewin and Knoll, 1999) and Belgium (Baele, 1999). Several occurrences of putative subsurface microbial fossils include salt dome cap rocks in the Texas Gulf (Sassen et al., 1988) and mineral veins in Germany (Hofmann, 1989; Walter and Reissmann, 1994) and Sweden (Pedersen et al., 1997). Evidence for former subsurface microbial activity has also been derived from corrosion patterns in marine basalts (Fisk et al., 1998; Furnes and Staudigel, 1999; Torsvik et al., 1998). Our initial report of lamentous structures in subsurface minerals (Hofmann and Farmer, 1997) was followed c 2000 Elsevier Science Ltd. All rights reserved. 0032-0633/00/$ - see front matter PII: S 0 0 3 2 - 0 6 3 3 ( 0 0 ) 0 0 0 8 1 - 7 1078 B.A. Hofmann, J.D. Farmer / Planetary and Space Science 48 (2000) 1077–1086 by a systematic search of museum collections and in the eld. Two main types of samples emerged from this study: Firstly, minerals (dominantly chalcedony and zeolites) from low-temperature hydrothermally altered volcanics, dominantly chalcedony and zeolites, and, secondly, samples from oxidized ore deposits. In both cases, mineralized laments formed a substrate for the nucleation and growth of paragenetically younger mineral phases. Small groups of samples are from solution cavities in limestones, serpentinized ultrabasic rocks and impactites. In total we have identied ¿140 occurrences in 65 geologically unrelated areas. Although tubular-shaped lamentous inclusions in chalcedony, of the type reviewed here, were previously described as fossil moss (Daubenton, 1782) and sponges (Bowerbank, 1842), inorganic interpretations have generally prevailed (Liesegang, 1915; Brown, 1957; Landmesser, 1984). However, since this early work, signicant advances have been made in our understanding of the processes of microbial taphonomy and fossilization through studies of ancient microbiotas (Golubic and Hofmann, 1976; Knoll, 1984) and their modern analogs (Cady and Farmer, 1996; Farmer and Des Marais, 1994; Ferris et al., 1988; Schultze-Lam et al., 1996). This work also provides a basis for re-evaluating the origin of lamentous and “stalactitic” minerals previously considered as inorganic precipitates. The fossilization of soft-bodied organisms occurs either through the alteration of original organic matter, or by its replacement with authigenic minerals (Allison and Briggs, 1991). Studies of microbial fossilization in rapidly mineralizing travertine and siliceous thermal spring systems indicate that this process more often involves the simple coating of external organic surfaces, with the rapid degradation of organic materials leaving behind external molds (Cady and Farmer, 1996). Often exterior organic surfaces, and the molds they produce, accumulate thin coatings of metalliferous precipitates which enhances their preservation during early diagenetic inlling and recrystallization. This process produces lamentous microfossils that have a tubular construction with an internal diameter similar to the size of the original bacterial lament (Fig. 1). Near deep sea vents, lamentous bacteria can become encrusted with silica, and the organic matter replaced by suldes or Fe-hydroxides (Zierenberg and Schiman, 1990; Duhig et al., 1992; Iizasa et al., 1998; Juniper and Fouquet, 1988; Little et al., 1999). In modern microbial mat organisms the diameter of lament molds produced by this process typically falls within the size range of 1–10 m. Variations in the inner core diameters of lament molds reect taxonomic or taphonomic dierences. Continued accretion and overgrowth and the merging of adjacent lament coatings can produce composite lament molds with outside diameters exceeding 100 m. Fig. 1. Subrecent siliceous sinter from Excelsior Geyser crater, Midway Geyser Basin, Yellowstone National Park, Wyoming, USA. Light microscope view showing encrusted laments of small heterotrophic organisms that resided within the framework of silicifying cyanobacterial (Calothrix) mats. In some examples, submicron molds are visible at the cores of opaline silica overgrowths where encrusted bacterial laments were later removed by oxidation, leaving behind external molds. Scale: Encrusted lament (with central mold) is ∼ 2 m in diameter. 2. Morphology and structure Filaments (Figs. 2 and 3) occur in high densities as inclusions in chalcedony, megaquartz, calcite, goethite, zeolites and some rarer minerals and are preserved due to encrustation by goethite, Fe-rich clay minerals or just by variations in grain size. Three lament morphotypes are tentatively recognized: (A) irregularly oriented laments, forming “moss agate” where enclosed in chalcedony; (B) laments showing gravity-induced vertical orientation, later overgrowths by minerals producing structures resembling stalactites; (C) strongly interconnected laments producing mat- or stromatolite-like fabrics of planar or rope-like geometry. In all cases examined, lamentous microstructures pre-date the formation of their host minerals as is evidenced by the occurrence of laments only thinly covered by minerals in open voids, and the orientation of mineral overgrowths showing radial growth of brous minerals nucleating on laments. Also, the persistence of laments across several nearby mineral grains, the restriction of lament inclusions to certain growth zones of host minerals and the presence of identical structures in dierent minerals like chalcedony and calcite from a single occurrence are evidence refuting the idea of a secondary emplacement (Feldmann et al., 1997). B.A. Hofmann, J.D. Farmer / Planetary and Space Science 48 (2000) 1077–1086 1079 Fig. 2. Macroscopic and microscopic appearance of subsurface lamentous fabrics occurring in goethite–chalcedony fracture inlls in Tertiary volcanic rocks, Sleeping Beauty Ridge, Cady Mts., California: (a) irregular goethite–chalcedony fracture with macroscopic laments bundles, scale = 10 cm; (b) hand-specimen (13 × 15 cm) showing streamer-like surface texture resulting from parallel orientation of mineralized laments; (c) bundles of laments encrusted with goethite and quartz showing vertical orientation due to gravity draping (height 5 cm). Though similar to stalactites, the lamentous interior shows that the stalks result from the encrustation of lament strands; (d) thin section of fracture inll showing lightly mineralized laments between heavily encrusted lament strands, embedded in brous quartz (chalcedony), height 2.0 mm; (e) thin section photomicrograph showing detail of mineralized laments in chalcedony, height 190 m. The occurrence of vertically oriented structures resulting from lament encrustation (morphotype (B) is widespread (Table 1). Such structures are commonly interpreted as stalactitic formations (e.g. Campbell and Barton, 1996). A nonstalactitic origin is indicated by the lamentous center close to 1 m in diameter, deviations from vertical alignment, and, in some cases, known subaqueous formation. Vertically arranged “pseudostalactitic” minerals are the most easily rec- ognized macroscopic expressions of mineralized laments of subsurface origin (Fig. 2c). In the subsurface examples documented here, laments have a tubular construction with inner core diameters of 0.2–8:5 m (average 1:5 ± 1:1 m, n = 486). This overlaps with the range of diameters observed in the tubular-shaped molds formed by lamentous bacteria in both modern subaerial thermal springs and deep-sea vents. In our 1080 B.A. Hofmann, J.D. Farmer / Planetary and Space Science 48 (2000) 1077–1086 Table 1 Selected occurrences of lamentous fabrics of subsurface origin and their main characteristics Locality Geological Age environment Filament mineralogy Late stage minerals Morphotypea Filament diameter (m) Howenegg, Hegau, Germany Iceland Faeroer Islands Cady Mts., California, USA Priday Ranch, Oregon, USA Needles, Texas, USA Sierra de Gallego, Chiuahua, Mexico Deccan Trapps, India Arz Bogd, Mongolia Parana basin, Brazil Campsie, Stirlingshire, Scotland Lake Superior area, USA Lengenbach, Valais, Switzerland Johanngeorgenstadt, Saxony, Germany Bescheert Gluck Mine, Saxony, Germany Herdorf near Siegen, Germany Huelgoat, Brittany, France Bleiberg, Carinthia, Austria Matagente, Cerro de Pasco, Peru Tsumeb, Namibia Broken Hill, New South Wales, Australia Hohenlimburg, Germany Warstein, Germany (Kretzschmar 1982) Ries Impact Crater, southern Germany V V V V V V V V V V V V OX OX OX OX OX OX OX OX OX SOL SOL I pyrite, opal goe,hem,cel cha goe clay minerals goe clay minerals cel,goe,hem,zeo clay minerals qtz hem hem goe goe goe goe goe goe goe goe goe goe goe ?, smectite zeo qtz,cha, zeo zeo cha,qtz,cc cha cha cha cha,qtz cha qtz cha cha void qtz cha goe Pb-phosphate goe,cha goe duftite goe qtz qtz,cc void C A,B,C A,B A,B,C A,B A A,B A,B A,C A,B A,B A,B C A,C A B C B B B A,B A,C A,C A Tertiary Tertiary Tertiary Tertiary Tertiary Tertiary Tertiary Cretaceous-Tertiary U. Jurassic–L. Cretaceous U. Jurassic–L. Cretaceous Devonian Precambrian (Proterozoic) Quarternary to Recent Quarternary to Recent Quarternary to Recent Quarternary to Recent Quarternary to Recent Quarternary to Recent Quarternary to Recent Quarternary to Recent Quarternary to Recent Tertiary Tertiary Tertiary 1:0 ± 0:3 (n = 21) 1:6 ± 0:9 (n = 70) 1:2 ± 0:4 (n = 9) 0:6 ± 0:2 (n = 29) 1:9 ± 0:7 (n = 29) 1:3 ± 0:4 (n = 7) 3:3 ± 0:5 (n = 9) 1:1 ± 0:4 (n = 20) 1:1 ± 0:4 (n = 18) 3:8 ± 1:9 (n = 25) 1:4 ± 0:3 (n = 17) 1:7 ± 0:5 (n = 11) 0:3 ± 0:1 (n = 14) 1:3 ± 0:5 (n = 22) 1:2 ± 0:3 (n = 23) 0:9 ± 0:5 (n = 7) 3:4 ± 0:9 (n = 14) 0:6 ± 0:2 (n = 10) 2:1 ± 0:7 (n = 13) 2:2 ± 0:6(n = 23) 0:8 ± 0:2 (n = 6) 0:6 ± 0:3 (n = 27) 1:1 ± 0:3 (n = 22) 0:5 ± 0:3 (n = 41) a Types: V = volcanic host rocks; OX = oxidized orebodies; SOL = solution cavities in limestone; I: impact melt rock. Mineral abbreviations: cc = calcite; cel = celadonite; cha = chalcedony; goe = goethite; hem = hematite; qtz = quartz; zeo = zeolites. Filament morphotypes: A: irregular (“moss agate”); B: gravity-induced orientation (pseudostalactites); C: mat=stromatolite fabrics. Fig. 3. SEM Micrograph showing goethite-encrusted laments from oxidized pyrite ore, Lengenbach, Binn Valley, Switzerland. Sample is from contact of orebody with overlying glacial till, about 5 m below surface. Scale bar is 100 m. subsurface examples, inner core diameters are fairly constant within any particular occurrence, although there is a certain size range between occurrences (Table 1). Compared with diameters of needle-shaped crystals with later mineral overgrowths that sometimes produce fabrics resembling mineralized laments, the frequency distribution of lament diameters is quite narrow (Fig. 4). Filament molds tend to become inlled during later mineral precipitation and the outer diameter of lamentous structures increases with continued accretion, nally reaching macroscopic dimensions. In modern microbial mats, the processes of growth, reproduction and taxis of lamentous species may lead to the formation of a variety of composite mat structures. In our subsurface examples, laments often occur in bundles ¡1–10 mm thick and up to 150 mm long. This lies within the size range of composite mat structures observed in living microbial mats. At the Cady Mts. (California) and Turnov (Czech Republic) sites, laments are bundled into mat-like structures of up to 10 mm diameter and 15 cm length, that resemble the “streamer” mats seen in areas of ow in modern thermal springs. Variable orientation of laments indicates that both gravity and water ow were factors controlling lament orientation. These composite structures are made up of tiny laments that were apparently only a few m in diameter and several cm long, indicating the former presence of exible, nonbrittle materials of high strength such as biopolymers. B.A. Hofmann, J.D. Farmer / Planetary and Space Science 48 (2000) 1077–1086 1081 Fig. 5. Twisted threads of goethite closely resembling modern Gallionella ferruginea embedded in chalcedony from cavity in Tertiary basalt, Breiddalur, Eastern Iceland. Scale bar = 50 m. and dierentiated forms interpreted to be fossil fungi or related groups have been reported previously from Warstein, Germany (Kretzschmar, 1982). At several sites in eastern Iceland we identied twisted stalks closely resembling the modern iron oxidizing bacterium Gallionella ferruginea (Fig. 5). 3. Environments of subsurface ÿlament formation Filamentous microstructures of probable microbial origin occur in numerous low-temperature subsurface environments that are typically characterized by fast mineral precipitation in large voids. The major types of geological environments are outlined briey below. 3.1. Volcanic host rocks Fig. 4. Histograms showing the frequency distribution of diameters of (A) lamentous microstructures and (B) of associated brous minerals of nonbiogenic origin. The laments from 21 dierent occurrences show a narrow peak close to 1 m, while nonbiogenic minerals have no preferred diameter. Filamentous structures from most investigated sites are branched (Figs. 2 and 3). In contrast to eucaryotes like the eumycota (fungi) and myxomyces (slime molds), individual bacterial laments do not truly branch during growth. However, during the early stages of fossilization of bacterial laments in modern mat systems, branched composite forms are often formed by coalescence of mineral overgrowths on adjacent laments. Dierentiating between true branching and coincidental joining of laments is dicult. In some of our samples, laments are morphologically differentiated, with attached globular structures 5 –70 m in diameter that resemble the sporangia of extant cellular slime molds (Madigan et al., 1997). Similar, complexly branched Amygdules (vesicles) and lithophysae as well as open fractures and ow-top breccias are the dominant forms of macroporosity which is commonly lled in by secondary minerals such as celadonite, Fe-hydroxides, zeolites, and various varieties of silica minerals including agate, chalcedony, megaquartz and opal. Filaments described here were often formed during an early stage of void inll. Energy sources to support microbial life in such environments include various types of redox couples, and possibly gases produced by water–rock interaction (Stevens and McKinley, 1995). The conditions during lament formation can only be deduced indirectly from the replacing minerals. Fluid inclusion homogenization temperatures measured in ◦ quartz cementing laments range from ¡70 to ¿200 C, indicating that late-stage mineral precipitation may have occurred at temperatures too high for life. The depth of formation can only be roughly estimated for most occurrences but is in the order of 10s to several 100 m. In the Cady Mts. (California) occurrence, 10 –50 cm-thick 1082 B.A. Hofmann, J.D. Farmer / Planetary and Space Science 48 (2000) 1077–1086 chalcedony veins containing goethite-encrusted laments crop out over a distance of 2 km, demonstrating the presence of signicant volumes of lament-rich rocks. Fig. 2 shows macroscopic mat-like fabrics, composed of lamentous forms, and microscopic lament architectures from this occurrence. 3.2. Oxidation zone of ore deposits Oxidizable ores such as suldes of Fe, Cu, Pb, Zn and Fe-rich carbonates are transformed into highly porous Fe-hydroxides minerals under near-surface conditions. Oxidation of sulde- and ferrous iron can serve as a source of energy for a number of autotrophic microbes (Nordstrom and Southam, 1997; Schrenk et al., 1998). Filamentous structures, often showing gravity-induced parallel orientation or mat-like coalescence of laments (Fig. 3) are very common in mineral samples from such environments and are commonly labeled as “stalactites” in collections. Formation temperatures were close to ambient, the depth of formation less than 10 to several 100 m (600 –1000 m in the case of Tsumeb, Namibia). 3.3. Solution cavities in carbonate rocks The rst Fe-encrusted laments of probable microbial origin were identied in low-temperature quartz deposited within solution cavities in massive Devonian limestone at Warstein, Germany (Kretzschmar, 1982). Similar occurrences are also known for several other sites within 50 km of Warstein. The age of these occurrences is poorly constrained, but is probably Tertiary. Filamentous micro organisms are enclosed in megaquartz and calcite. Quartz crystals are of the “Suttrop-Type” (Behr et al., 1979; Behr and Horn, 1984) and record a complex uid history, with laments being present only in the youngest growth zones. Fluid inclusion homogenization temperatures for the entombing quartz ◦ indicate crystallization at 100 –160 C after lament formation (Kretzschmar, 1982; this study). 3.4. Open space in macrofossils Pyritized ammonites from Jurassic black shales in England and Germany (Hudson, 1982) and from the Jurassic of Switzerland and the Cretaceous of southern France (own observations) contain geopetal features interpreted as stalactites (Hudson, 1982). Based on the similarity with pseudostalactitic features from volcanic and oxidation zone environments an interpretation as pyrite-encrusted, gravityoriented microbial laments appears more likely. 3.5. Other environments A small number of samples showing structures resembling lamentous biofabrics are derived from hydro- thermal veins in crystalline rocks (Campbell and Barton, 1996; Hofmann, 1989; Walter and Reissmann, 1994), Mississippi-Valley-type Pb–Zn-deposits (Peck, 1979), chalcedony veins in serpentinites (NMBE collection) and impact melt rocks (Ries, Germany, NMBE collection). Subaqueously formed carbonate speleothems (“pool ngers”) that appear to have formed by calcication of microbial laments (Adolphe et al., 1991; Davis et al., 1990) are the closest recent analogue to lamentous subsurface biofabrics of pseudostalactitic habit. 4. Mineralogy and geochemistry The lamentous structures studied here are systematically associated with iron minerals. Fe-rich clays and pyrite dominate in reduced mineral assemblages, while goethite and hematite dominate oxidized assemblages. The commonly observed replacement of goethite by Fe-rich clays indicates that goethite was a very common primary mineral on the laments, perhaps even where it is now completely altered. The mineralogy of the laments is relatively monotonous. Three phases of mineralization may be distinguished: (A) Initial precipitation forming laments of approximately 1 m diameter and composites of laments. Minerals precipitated at this stage are ne grained and do not show preferential orientation. Fe hydroxides, celadonite and pyrite are the most common phases. (B) Deposition of later minerals on laments, typically increasing the lament diameter to 20 –40 m. Minerals are radially oriented around tubular laments and include chalcedony, goethite, hematite, clay minerals, zeolites and a wide variety of oxidation zone species. (C) Final cementation, leading to stalk-like forms up to several cm in diameter and several tens of cm long. The mineralogy is dominated by quartz varieties and= or calcite (hydrothermal) or a large variety of oxidation stage minerals (oxidation zones). Where stage (C) cementation is not complete, mineral-coated laments are present in open pore space (Fig. 3). An important observation is the fact that nearly identical lament geometries are present in samples of variable mineralogy, both in a single deposit and in distant occurrences. This indicates that the occurrence and morphology of mineralized laments is independent of the host mineralogy, and is unlikely to be the product of specic mineral forming reactions. The abundance of goethite as an early mineral in lament mineralization is consistent with formation ◦ temperatures ¡100 C (Diakonov et al., 1994). The B.A. Hofmann, J.D. Farmer / Planetary and Space Science 48 (2000) 1077–1086 1083 6. Discussion Probably the strongest argument for biogenicity is the tubular architecture of the laments at a constant diameter, and the coalescence of these laments into mat-like structures, often showing gravity draping. While various lamentous and spiral forms can be produced by growth in gels, we are not aware of any nonbiological process producing tubular laments of this narrow size range (Figs. 2 and 4). Gravity draping (Fig. 2) is very common and suggests that the laments initially consisted of exible, nonrigid material such as an organic polymer. Other features supporting a biological interpretation are the branched morphology (Figs. 2 and 3) and the presence of dierentiated structures, such as twisted stalks (Fig. 5) and globules associated with some laments. Finally, laments are typically present in mineral assemblages formed at tem◦ peratures below ∼100 C based on the abundance of primary goethite and geological context (Table 1). We nd the similarity between these features and those observed in fossilized materials of modern terrestrial (Fig. 1; Cady and Farmer, 1996) and submarine (Pracejus and Halbach, 1996; Zierenberg and Schiman, 1990) thermal spring environments to be especially compelling. The lack of organic carbon in our structures is not surprising, considering that they originate on microbial laments of approximately 1 m diameter that contained just a few weight percent organic carbon, and the rather oxidizing environment indicated by abundant goethite. Among possible abiogenic origins of laments (brous crystals, speleothems, uid inclusion trails, vermiform crystal aggregates), none is compatible with gravity-draped tubular laments of very restricted size range. Nonbiological explanations such as “biomorphs”, noncrystallographic mineral precipitates experimentally formed in silica-rich environments (Garcia-Ruiz, 1981, 1994; Garcia-Ruiz and Amoros, 1998) cannot explain the very restricted size range in our samples, the independence of lament size and fabric architecture from the type of environment, and the gravity draping. 6.1. Arguments supporting a biogenic origin of ÿlamentous subsurface fabrics 6.2. Involvement of microbes in mineral precipitation The subsurface microstructures described here most likely were formed by fossilization of lamentous microorganisms. We infer a biological origin for the lamentous forms found in our subsurface examples, based on comparisons with the style of microbial preservation reported from modern mat systems found in rapidly mineralizing thermal spring environments (Fig. 1). Both fossil lamentous microbes from thermal springs and our subsurface laments consist of curved, tubular and branching mineral laments with constant core diameters, are largely devoid of organic matter and closely resemble known biological structures. The glycoproteins of bacterial cell walls have been shown to passively bind metal ions, and some metals (e.g. Fe, Mn, Ni, and V) are known to be important components of metalloenzymes and required co-factors for cell growth (Ehrlich, 1986). Thus, these elements are actively concentrated by many heterotrophic micro-organisms, which enhances the opportunity for early diagenetic mineralization. The precipitation of Fe- and Mn-oxides may be directly mediated by physiological processes, with ne-grained metal oxides forming either intracellularly, e.g. as in some magnetotactic bacteria (Frankel et al., 1983), or accumulating on cell surfaces (Ferris homogenization temperatures of uid inclusions in quartz ◦ postdating lament formation ranges from ¡70 to 160 C for several occurrences, indicating a rise in temperature after lament formation. Filaments do not contain microscopically visible remnant organic matter, which is consistent with the oxidation of organic materials shortly after entombment. This style of preservation was also noted in both modern (Cady and Farmer, 1996) and Devonian-aged siliceous sinters from Queensland (Walter et al., 1996b; Walter et al., 1998). 5. Estimating the abundance of subsurface biofabrics The probability to nd subsurface biofabrics in a randomly taken sample from volcanic rocks must be considered low because secondary minerals constitute only a very small fraction of such rocks. However, among samples of secondary mineral precipitates from volcanics and oxidized ores, subsurface biofabrics are quite common. A rough estimate can be made based on the abundance in mineral collections. Among 209 investigated specimens of quartz and zeolites from volcanic rocks and of oxidation zone minerals in the collections of the Bern Natural History Museum, 14% of samples from volcanic environments and 21% of samples from oxidized ore deposits contained lamentous structures of probable microbial origin (only samples collected prior to this study counted). This estimate may be biased for a number of reasons with possible under- and overrepresentation of lamentcontaining samples. Still, the common occurrence of subsurface biofabrics is demonstrated to such an extent that they can be regarded as an integral constituent of volcanic formations having undergone hydrous low-T alteration, and of oxidized ore deposits. 1084 B.A. Hofmann, J.D. Farmer / Planetary and Space Science 48 (2000) 1077–1086 et al., 1988) or within exopolymer matrices (Geesy and Jang, 1989). The initial precipitation of Fe-hydroxides as an early phase could be due to active physiological mediation of Feprecipitation by organisms. However, studies of modern thermal springs suggest that it is more likely that the bulk of the precipitation is driven by inorganic processes such as changes of temperature, pH and redox potential. Organisms, nevertheless, provide surfaces for the initial nucleation of minerals, in turn setting up morphological patterns that persist during subsequent growth and accretion. While this can result in a lack of preserved isotopic or other geochemical signatures for life, it may also lead to the formation of characteristic micro- and mesoscale biofabrics that are indicative of life (Farmer and Des Marais, 1994). 6.3. Types of involved micro-organisms The types of micro-organisms potentially responsible for the formation of the subsurface lamentous structures discussed here is uncertain. The size range of the laments and inferred preservational environments are consistent with a bacterial, archeal, or fungal origin. Energy requirements of micro organisms under such conditions could have been met by redox couples such as O2 –H2 S, O2 – Fe2+ or H2 – CO2 . However, the simple lamentous forms described here do not necessarily require biochemical simplicity. In fact, most heterotrophic communities contain many smaller lamentous and coccoidal species that are not easily preserved. Thus, taphonomic biases may mask the full range of metabolic complexity possible within microbial consortia (Jorgensen et al., 1992). Iron oxidizers are indicated as possible key candidates based on morphology (Fig. 5) and the predominance of ferric iron minerals as early precipitates on laments. 6.4. Application in the search for fossil life on Mars Based on the currently available information on the geology of Mars it appears very likely that formerly watersaturated macroporous subsurface environments (e.g. vesicular and brecciated volcanic and impact rocks) are very common and may even constitute major fractions of the upper Martian crust. Sulde-rich rocks and iron carbonates, possible energy sources in presence of an oxidant, may also be present based on evidence from Martian meteorites. Rocks very similar to those hosting widespread terrestrial subsurface microbial fabrics are thus most likely present on Mars. It appears straightforward, therefore, to use terrestrial subsurface microbial fabrics as an analogue for potential remnants of similar Martian life forms. Potential target areas include all features indicative of possible water percolation (fractures, vesicles, secondary minerals). The macroscopic characteristics of lamentous fabrics (e.g. pseudostalactites) would provide means for relatively easy recognition. Unambiguous interpretation of possible lamentous structures recognized on Mars will depend heavily on determining the modes of origin of terrestrial counterparts. 6.5. Further research on subsurface ÿlaments Further research will be needed to provide additional evidence for the proposed biogenicity of the subsurface laments described in this study. Arguments to be tested include the presence of organic carbon in the center of laments, enrichments of trace elements in lament cores, and fractionation of stable isotopes (e.g. Fe). A more detailed characterization of the environments of formation using petrography, stable isotopes and uid inclusions is also planned. This will help to identify possible sites of active lament formation. Such sites will be needed to investigate the process of the actively ongoing lament mineralization in a way similar to studies in mineralizing hot spring environments (Farmer and Des Marais, 1994; Cady and Farmer, 1996). 7. Conclusions Filamentous fabrics of possible microbial origin are common in terrestrial subsurface environments. Even though only a small fraction of subsurface biota may be fossilized in such a way, these micro- to macroscopic scale features are interpreted to be morphological evidence for the widespread occurrence of subsurface life in the geological past. Mineralization of microbes appears to be a common process in low-temperature subsurface environments. Unexplained morphological features such as certain “stalactites” with micron-sized cores (Campbell and Barton, 1996) and nondendritic moss agates may have been formed by the encrustation and replacement of microbial laments and associated composite mat features within environments previously considered to be devoid of microbial remains. Our results open perspectives for exploring new aspects of the fossil record of life on Earth and possibly on Mars and other planetary bodies. On Mars, subsurface environments consisting of porous volcanic rocks with water likely existed in the past and may even persist today as deep aquifers conned beneath a capping cryosphere of shallow crustal ground ice (Cliord, 1993). Modern and ancient groundwater environments have been identied as important targets for future exploration and sample return by the U.S. Mars Global Surveyor Program (McCleese, 1996). Our study suggests a potentially productive line of inquiry in the exploration for a Martian fossil record. Volcanic terranes are widespread on Mars and are likely to host a variety of low-temperature hydrothermal deposits and secondary minerals disseminated though volcanic host rocks as fracture (vein) and amygdule lls. Such secondary minerals are likely to be present in abundances too low to be detected by orbital remote sensing and may require the use of instrumented rovers for detection. B.A. Hofmann, J.D. Farmer / Planetary and Space Science 48 (2000) 1077–1086 Acknowledgements We thank all those who provided samples for the present study, particularly J. Arnoth, M. Helfer, A. Klee, P. and G. Penkert. We thank P. Vollenweider for photographic work and Ph. Hauselmann for uid inclusion measurements. Work of BAH was partially funded by the Swiss National Science Foundation and Entwicklungsfonds Seltene Metalle (Pully, Switzerland). Work of JDF was funded by grants from NASA’s Astrobiology Institute, the NASA Exobiology Program and the Director’s Discretionary Fund, NASA Ames Research Center. References Adolphe, J.-P., Choppy, B., Loubiere, J.-F., Paradas, J., Loleilhavoup, F., 1991. Biologie et concretionnement: un example, les baguettes de gours. Karstologia 18, 49–55. Allison, P.A., Briggs, D.E.G., 1991. Taphonomy of soft-bodied animals. In: Donovan, S.K. (Ed.), The Processes of Fossilization. Columbia University Press, New York, pp. 120–140. Baele, J.-M., 1999. Karst and silication in continental and margino-littoral environment: two examples in the Meso-Cenozoic of Southern Belgium. In: Audra, P. (Ed.), Karst 99, Grand Causses, Vercors, France. Universite de Provence, pp. 49 –54. Behr, H., Hess, H., Oehlschlegel, G., Lindenberg, H.G., 1979. Die Quarzmineralisation vom Typ Suttrop am N-Rand des rechtsrheinischen Schiefergebirges. In: Meiburg, P. (Ed.), Geologie und Mineralogie des Warsteiner Raumes. Der Aufschluss, Sonderband 29, Heidelberg, pp. 205 –231. Behr, H.J., Horn, E.-E., 1984. Quarzbildung und Verkieselungsprozesse in den Karbonatkomplexen des Rheinischen Schiefergebirges. In: Walter, H.W. (Ed.), Postvariszische Gangmineralisationen in Mitteleuropa. Verlag Chemie, Weinheim, pp. 27–45. Bowerbank, J.S., 1842. On the spongeous origin of moss agates, other siliceous bodies. AMNH 10, 9 –18, 84–91. Brown, R.W., 1957. Plantlike features in thunder eggs and geodes. Annual Report of the Board of Regents of the Smithsonian Institution, pp. 329 –339. Cady, S.L., Farmer, J.D., 1996. Fossilization processes in siliscours thermal springs: trends in preservation along the thermal gradient. In: Bockand, G.R., Goode, J.A. (Eds.), Evolution of Hydrothermal Ecosystems on Earth (and Mars?). Wiley, Chichester, pp. 150–173. Campbell, W.R., Barton, P.B., 1996. Occurrence and signicance of stalactites within the epithermal deposits at Creede. Colorado. Can. Mineral. 34, 905–930. Cliord, S.M., 1993. A model for the hydrologic and climatic behavior of water on Mars. J. Geophys. Res. 88, 2456–2474. Dahanayake, K., Krumbein, W.E., 1986. Microbial structures in oolitic iron formations. Miner. Deposita 21, 85–94. Daubenton, L.J.M., 1782. Sur les causes qui produisent trois sortes d’herborisations dans les pierres. Mem. Acad. Roy. 667– 673. Davis, D.G., Palmer, M.V., Palmer, A.N., 1990. Extraordinary Subaqueous speleothems in Lechuguilla Cave, New Mexico. NSS Bull. 52, 70–86. Delaney, J.R., Kelley, D.S., Lilley, M.D., Buttereld, D.A., Baross, J.A., Wilcock, W.S.D., Embley, R.W., Summit, M., 1998. The quantum event of oceanic crustal accrection: impacts of diking at mid-ocead ridges. Science 281, 222–230. Dexter-Dyer, B., Kretzschmar, M., Krumbein, W.E., 1984. Possible microbial pathways in the formation of Precambrian ore deposits. J. Geol. Soc. (London) 141, 251–262. Diakonov, I., Khodakovsky, I., Schott, J., Sergeeva, E., 1994. Thermodynamic properties of iron oxides and hydroxides. I. Surface 1085 and bulk thermodynamic properties of goethite (a-FeOOH) up to 500 K. Eur. J. Mineral. 6, 967–983. Duhig, N.C., Davidson, G.J., Stolz, J., 1992. Microbial involvement in the formation of Cambrian sea-oor silica–iron oxide deposits, Australia. Geology 20, 511–514. Ehrlich, H.L., 1986. Geomicrobiology. Marcel Dekker, New York. Farmer, J.D., Des Marais, D.J., 1994. Biological versus inorganic processes in stromatolite morphogenesis: observations from mineralizing sedimentary systems. In: Stal, L.J., Caumette, P. (Eds.), Microbial Mats: structure development and environmental signicance. NATO ASI series, G35, Springer, Berlin. pp. 61–68. Farmer, J.D., Des Marais, D.J., 1999. Exploring for a record of ancient Martian life. J. Geophys. Res. 104 (E-11), 26 977–26 995. Feldmann, M., Neher, J., Jung, W., Graf, F., 1997. Fungal quartz weathering and iron crystallite formation in an Alpine environment, Piz Alv, Switzerland. Eclogae Geol. Helv. 90, 541–556. Ferris, F.G., Fyfe, W.S., Beveridge, T.J., 1988. Metallic ion binding by Bacillus subtilis: implication for the fossilization of microorganisms. Geology 16, 149–152. Fisk, M.R., Giovannoni, S.J., Thorset, I.H., 1998. Alteration of oceanic volcanic glass: textural evidence of microbial activity. Science 281, 978–980. Frankel, R.B., Papaefthymiou, G.C., Blakemore, R.P., O’Brien, W., 1983. Fe3 O4 precipitation in magnetotactic bacteria. Biochim. Biophys. Acta. 763, 147–159. Furnes, H., Staudigel, H., 1999. Biological mediation in ocean crust alteration: how deep is the deep biosphere?. EPSL 166, 97–103. Garcia-Ruiz, J.M., 1981. Crystal aggregates with induced morphologies grown by silica gel technique. Bull. Mineral. 104, 107–113. Garcia-Ruiz, J.M., 1994. Inorganic self-organization in Precambrian cherts. Origins Life Evol. Biosphere 24, 451–467. Garcia-Ruiz, J.M., Amoros, J.L., 1998. Carbonate precipitation into alkaline silica-rich environments. Geology 26, 843–846. Geesy, G.G., Jang, L., 1989. Interactions between metal ions and capsular polymers. In: Beveridge, T.J., Doyle, R.J. (Eds.), Metal Ions and Bacteria. Wiley, New York, pp. 325–357. Ghiorse, W.C., 1997. Subterranean life. Science 275, 789–790. Gold, T., 1992. The deep, hot biosphere. Proc. Natl. Acad. Sci. USA 89, 6045–6049. Gold, T., 1999. The Deep Hot Biosphere. Copernicus, New York, 235pp. Golubic, S., Hofmann, H.J., 1976. Comparison of Holocene and mid-Precambrian Entophysalidacea (Cyanophyta) in stromatolitic algal mats: cell division and degradation. J. Paleontol. 50, 1074–1082. Hofmann, B., 1989. Genese, Alteration und rezentes Fliess-System der Uranlagerstatte Krunkelbach (Menzenschwand, Sudschwarzwald), Nagra Technical Report, Baden, Switzerland. Hofmann, B.A., Farmer, J.D., 1997. Microbial fossils from terrestrial subsurface hydrothermal environments: Examples and implications for Mars. In: Cliord, S.M., Treiman, A.H., Newsom, H.E., Farmer, J.D. (Eds.), Conference on Early Mars: Geologic and Hydrologic Evolution, Physical and Chemical Environments, and the Implications for Life. Lunar and Planetary Science Institute, Houston, pp. 40 – 41. Hudson, J.D., 1982. Pyrite in ammonite-bearing shales from the Jurassic of England and Germany. Sedimentology 29, 639–667. Iizasa, K., Kawasaki, K., Maeda, K., Matsumoto, T., Saito, N., Hirai, K., 1998. Hydrothermal sulde-bearing Fe–Si oxyhydroxide deposits from the Coriolis Troughs, Vanuatu backarc, southwestern Pacic. Mar. Geol. 145, 1–21. Jorgensen, B.B., Nelson, D.C., Ward, D.M., 1992. Chemotrophy and decomposition in modern microbial mats. In: Schopf, J.W., Klein, C.C. (Eds.), The Proterozoic Biosphere: A Multidisciplinary Study. Cambridge University Press, Cambridge, pp. 287–293. Juniper, S.K., Fouquet, Y., 1988. Filamentous iron–silica deposits from modern and ancient hydrothermal sites. Can. Mineral. 26, 859–869. Knoll, A.H., 1984. Exceptional preservation of photosynthetic organisms in silicied carbonates and silicied peats. Philos. Trans. Roy. Soc. (London) 311B, 111–122. 1086 B.A. Hofmann, J.D. Farmer / Planetary and Space Science 48 (2000) 1077–1086 Kretzschmar, M., 1982. Fossile Pilze in Eisen-Stromatolithen von Warstein (Rheinisches Schiefergebirge). Facies 7, 237–260. Landmesser, M., 1984. Das Problem der Achatgenese. Pollichia 72, 5–137. Liesegang, R.E., 1915. Die Achate. Theodor Steinkop, Dresden und Leipzig. Little, C.T.S., Herrington, R.J., Haymon, R.M., Danelian, T., 1999. Early Jurassic hydrothermal vent community from the Franciscan Complex, San Rafael Moutains, California. Geology 27, 167–170. Madigan, M.T., Martinko, J.M., Parker, J., 1997. Biology of Microorganisms. Prentice-Hall, London, 986pp. McCleese, 1996. The Search for Evidence of Life on Mars. NASA (www.hq.nasa.gov=oce=oss=mccleese.htm). McKay, D.S., Gibson, E.K., Thomas-Keprta, K.L., Vali, H., Romanek, C.S., Clemett, S.J., Chillier, X.D.F., Maechling, C.R., Zare, R.N., 1996. Search for past life on Mars: possible relic biogenic activity in Martian meteorite ALH84001. Science 273, 924–930. Nordstrom, D.K., Southam, G., 1997. Geomicrobiology of sulde mineral oxidation. In: Baneld, J.F., Nealson, K.H. (Eds.), Geomicrobiology, Reviews in minerology 35. Mineralogical Society of America, Washington, D.C. pp. 361–390. Parkes, R.J., Cragg, B.A., Bale, S.J., Getli, J.M., Goodman, K., Rochelle, P.A., Fry, J.C., Weightman, A.J., Harvey, S.M., 1994. Deep bacterial biosphere in Pacic Ocean sediments. Nature 371, 410–413. Peck, S.B., 1979. Stalactites and helictites of marcasite, galena, and Sphalerite in Illinois and Wisconsin. NSS Bull. 41, 27–30. Pedersen, K., Ekendahl, S., Tullborg, E., Furnes, H., Thorset, I., Tumyr, O., 1997. Evidence of ancient life at 207 m depth in a granitic aquifer. Geology 10, 827–830. Pracejus, B., Halbach, P., 1996. Do marine moulds inuence Hg and Si precipitation in the hydrothermal JASDE eld (Okinawa Trough). Chem. Geol. 130, 87–99. Sassen, R., McCabe, C., Kyle, R., Chinn, E.W., 1988. Deposition of magnetic pyrrhotite during alteration of crude oil and reduction of sulfate. Org. Geochem. 14, 381–392. Schrenk, M.O., Edwards, K.J., Goodman, R.M., Hamers, R.J., Baneld, J.F., 1998. Distribution of Thiobacillus ferrooxidans and Leptosprillium ferrooxidans: implications for generation of acid mine drainage. Science 279, 1519–1522. Schultze-Lam, S., Fortin, D., Davis, B.S., Beveridge, T.J., 1996. Mineralisation of bacterial surfaces. Chem. Geol. 132, 171–181. Stevens, T.O., 1997. Subsurface microbiology and the evolution of the biosphere. In: Penny, A., Haldeman, D. (Eds.), The Microbiology of the Terrestrial Deep Subsurface. Lewis Publishers, Boca Raton, NY, pp. 205–223. Stevens, T.O., McKinley, J.P., 1995. Lithoautotrophic microbial ecosystems in deep basalt aquifers. Science 270, 450–454. Torsvik, T., Furnes, H., Muehlenbachs, K., Thorseth, I.H., Tumyr, O., 1998. Evidence for microbial activity at the glass-alteration interface in oceanic basalts. EPSL 162, 165–176. Trewin, N.H., Knoll, A.H., 1999. Preservation of Devonian chemotrophic lamentous bacteria in calcite veins. Palaios 14, 288–294. Walter, H., Reissmann, R., 1994. Organische(?) Strukturen in Achatgangen dews Osterzgebirges. Palaont. Z. 68, 5–16. Walter, M.R., 1996a. Ancient hydrothermal ecosystems on earth: a new palaeobiological frontier. In: Bock, G.R., Goode, J.A. (Eds.), Evolution of Hydrothermal Ecosystems on Earth (and Mars?). Wiley, Chichester, pp. 112–127. Walter, M.R., Des Marais, D.J., Farmer, J.D., Hinman, N.W., 1996b. Paleobiology of mid-Paleozoic thermal spring deposits in the Drummond Basin, Queensland, Australia. Palaios 11, 497–518. Walter, M.R., Mcloughlin, S., Drinnan, A.N., Farmer, J.D., 1998. Paleontology of Devonian thermal spring deposits, Drummond Basin, Australia. Alcheringa 22, 285–314. Zierenberg, R.A., Schiman, P., 1990. Microbial control of silver mineralization at a sea-oor hydrothermal site on the northern Gorda Ridge. Nature 348, 155–157.